My top ten predictions for the data center industry in 2014. In order from “extremely likely” to “I’m totally guessing here”.
Okay. Let’s dispense with a few sure bets first,
1. Big Data will continue to be a challenge and an opportunity.
The past few years have seen dramatic growth in the amount of raw data produced and stored by corporations, individuals, governments and scientists. IBM estimates that a staggering 2.5 quintillion bytes of data are created each day around the world. All that data obviously has to be stored somewhere. Which points toward a continuing demand for most types of data center space and data storage devices.
For corporations, it’s not unusual for a single US business to have more than 100,000 gigabytes of stored data. Warehousing these vast stores of data is relatively straightforward and becoming easier and cheaper as the cost of memory nears the vanishing point. The challenge lies in putting that data to work. The relational databases and desktop applications that processed and helped visualize data in the past are increasingly unsuitable for dealing with the scale of modern data hoards.
Simply analyzing massive volumes of data is a true challenge. But the problem goes deeper. The data is also arriving at increasing velocity and in an increasing variety of formats. Making good strategic use of Big Data requires that the analytics tools be capable of analyzing the vast volume of data in real time and across a variety of data types.
In 2013 investment in Big Data neared $1.4B USD. In 2014, the companies that bring analytics tools to market that successfully capture, analyze and visualize Big Data will see increased investment and demand for services.
2. Cloud Computing will continue to grow.
All types of cloud computing solutions will see strong growth in 2014. However, growth in the area of cloud based data storage will be particularly strong. Market research firm ABI forecasts that cloud based data storage will triple in volume in the next 5 years. ABI is expecting 4 billion personal accounts holding a whopping 3500 petabytes of data to be in place by 2018.
Further driving growth in cloud computing will be an increasing trend among enterprises, SMBs and startups to decide against owning and maintaining their own stack of silicon. Business will see that the costs associated with owning IT infrastructure as an impediment to the pursuit of their core mission. These firms will continue the rapidly adopt PaaS, SaaS and HaaS cloud solutions.
3. Colocation and managed hosting will continue to grow.
Perceived security and reliability concerns will cause many companies to remain reluctant to make a full commitment to the cloud. However, these companies will still see the logic in deciding to forgo the expense of building, operating and maintaining the infrastructure needed for their own private data center. These companies will turn to hybrid cloud, colocation and hosting providers for their technology backbone.
Growth for colos in 2014 should be strong in all geographies. For example, a recent study by Research and Markets indicates an expected 16.5% CAGR for European colocation providers through 2016.
4. Data Center Infrastructure Management (DCIM) Market to Grow
Data center operators will continue to look for strategies that allow them to squeeze as much efficiency out of their existing infrastructure as possible. Maximizing the use of existing electrical, mechanical and compute infrastructure requires careful planning, thoughtful deployment and a clear view of available resources. A number of DCIM technology platforms have demonstrated significant capability provide this vision and planning capability. As a result, the DCIM market has exploded in the past two years
In December, Gartner, an information technology research and advisory firm, stated that the market for DCIM products is already to more than 1B$. Look for DCIM to continue its upward trend in 2014.
A little less obvious…
5. A single provider for Colo, Cloud and Big Data Analytics
These first 3 predictions comprise what I see as a winning business model for the IT infrastructure outsourcing firm of the future. Companies will emerge that wrap all of these services together into a tidy package and own their customer’s IT relationship from physical layer through advanced data analytics.
A market leader will emerge that provides:
- Colocation and managed services for those customers seeking a conservative approach to IT infrastructure outsourcing
- Help those customers gradually transition to a hybrid cloud solution
- Help customers transition from hybrid cloud to a fully functional cloud solution
- Provide in-house Big Data analysis and consulting
Will companies like AWS that provide cloud and big data analytics move down the technology ladder to provide traditional colocation services? Will companies like Equinix that provide colocation and cloud services move up the technology ladder to provide big data analytics? 2014 should tell us.
6. The Internet of Things (IoT) inches closer
As previously noted, Big Data will continue to be a focus for technology firms in 2014 and beyond. Perhaps the biggest data hoard that this tech will be tasked with taming is the deluge of data poised to arrive when nearly every object in the world is part of the Internet of Things and is actively recording and reporting data.
Gartner estimates that by 2020, 26 billion devices will be on the IoT.
The IoT will inch closer in 2014 with two key conferences:
These conferences will incubate ideas, demonstrate capabilities and establish common protocols for IoT tech.
As noted above, cascades of data lead directly to demand for data center space and boom times for data center designers and infrastructure providers.
7. Rise of the Robots!
According to the Uptime Institute’s Abnormal Incident Reporting (AIR) database, human error accounts for 70% of unplanned data center downtime. (As Uptime’s Hank Seader points out, at root cause, nearly ALL data center failures can be attributed to human error.)
There are some strategies that will help reduce these errors. For example;
- As Schneider Electric points out, knowledgeable, highly trained data center operators using tested, formal operating procedures can go a long way toward reducing unplanned downtime.
- DCIM tools that provide actionable, accurate and timely data regarding infrastructure conditions can also decrease human error.
But if you really want to start reducing data center downtime due to human error we need to consider replacing data center humans with data center robots.
We have already starting to see a few robots creep into data center electrical rooms. Remote circuit breaker racking solutions such as those offered by CBS ArcSafe are technically robots.
But more significantly, in 2013 Google quietly purchased 8 robotics companies including the industry leader, Boston Dynamics. Boston Dynamics created the spooky Big Dog and Cheetah robots and the DARPA competition ATLAS robots.
Google has been tight-lipped regarding its plans for their newly acquired robot tech. But I’m betting that the data centers on Google’s drawing board will look more like the robot controlled library at the University of Chicago or the inside of a tape drive silo.
Imagine robots racking and stacking servers in 50’ tall racks.
8. Data Center Energy efficiency shifts focus to the left side of the PUE decimal place
Ever since the EPA dropped its report to Congress in 2009, a steady stream of green metrics, technologies and strategies have emerged. Most of this tech is focused on reducing the power consumption associated with the mechanical, electrical and other ancillary systems that data centers require. Incredible feats have been accomplished and energy use has been reduced greatly. As a result, many new data centers are seeing legitimate <1.1 PUEs.
However, many late model data centers are tapped out when it comes to mechanical and electrical system energy efficiency. It’s time for the IT manufacturers to start making reliable, innovative leaps forward in server efficiency. In 2014, I expect to see companies like Servergy lead a march into a new era of data center efficiency.
Finally, some harsh truths for a couple of groups that are not likely see growth…
9. Local and regional manufacturer’s rep firms feel the pinch.
The increasing number of small and medium businesses that decide against building their own data centers and server rooms is bad news for the local manufacturer’s representatives. The guys that sell racks, power strips, servers, UPS systems and other infrastructure to this marketplace are in for a bumpy ride as their marketplace rapidly shrinks.
Manufacturer’s reps will need to move upmarket and fight it out for the business of the growing colos and cloud providers if they hope to survive. Unfortunately, the manufacturer rep firms will find that the colo and cloud owners are extremely adept at finding the margin in infrastructure deals and cutting it out.
As a result, most of the colo and cloud accounts are being taken direct by the manufacturers who are also struggling in an increasingly savvy and competitive marketplace.
10. Fewer data center facilities personnel needed
Any time an industry adopts a technology that radically increases efficiency and automates tasks that were previously done by labor, jobs disappear. In fact, the whole point behind developing efficiency and automation is to reduce the amount of labor needed. Reduced labor is another term for fewer jobs.
DCIM, robots and other automation tools are allowing data centers to squeeze efficiency from their infrastructure and will eventually allow them to trim their labor force.
If a significant part of your job consists of walking around a data center with a clipboard and taking readings from data center infrastructure GUIs, somewhere someone is pitching the idea that a DCIM sensor and software package can replace you.
The sheer number of data centers that are needed to house all of our data should result in a net gain in data center jobs. However, the days when data center personnel could be “facilities” only are rapidly drawing to close. To be competitive in the 2014, workers must understand the infrastructure, the software that monitors and controls it and the mission critical loads it supports.
That’s that I think. What do you think 2014 holds for us?